Today I’m going to a local mini-conference to discuss “Social Graph Unification”, a topic that’s been hot the past few weeks. It’s exciting to think that you could unify your friend lists across Facebook, MySpace, and Gmail. But we should not be content with a mere synchronizing of friend lists. Friend lists are merely one class of data that you might like to have synchronized. (In fact, I think it’s stupid to have your replicated across different services, but that’s a post for a different day!) Here is a rough sketch of the different types of data that you might create online these days, across all services. The question for today is, how many of these would you like (or be comfortable with) being synchronized?
| Profile | Identities | Relationships | Content | Communication |
| Data about the user themselves, rarely changed. | Other online identities of this user. Assertions of identity. | Expressions of trust, interest,etc.. towards other users. Uni-or bi-directional. | Data created by the user, available for others. Serialized. ‘Pull’ | Data created by the user, directed at 1 other (or selected few) ‘Push’ |
| Name Photo Age/Birthday Gender Location Relationship Status Interests Political Views Religious Views School History Job History Favorite Music/Books/Quotes Many more… |
Email address Homepage Many more… |
Friends Contacts Subscriptions
|
Status Blog Posts Photos Videos Reviews Bookmarks Stuff For Sale Wall Posts* Group/Forum Posts Clickstream** Searchstream** |
Email IM Phone/Video Calls |
Interesting dimensions to think about:
Pointers vs. Raw Data: Identities and Relationships are not data themselves, but point to other data sources. Profile and Content are data in themselves. Bookmarks are a type of content that is a pointer to other content.
Serialized vs. In-Place: Profile data is typically not serialized, but Content data is. Some fields are in flux between them: E.g. Twitter showed that “Status” could be serialized, and other companies are trying to serialize “Location” (So that you could a feed of someone’s location history.)
Push vs. Pull: Blog posts are messages that are put out there for anyone to ‘pull down’ are read. Emails are messages ‘pushed’ directly to someone else.
Also interesting to think of the coverage that different services achieve. General purpose social networks like MySpace and Facebook cover almost all most of these types of data. Other services have carved out specific niches: LinkedIn does a little bit of profile data, business contacts, and email. Twitter does a little bit of profile information, subscriptions, and status. Flickr does a little bit of profile information, subscriptions, friends, photos, and email.
Will people remain content to have this data scattered across services? Will syncronization enable this? How the hell do you manage permissions across all of this!? I’ll let you know how the conference goes!
Great breakdown.. I always learn something 🙂
Nice. Thanks for putting this up. It’s a nice introduction to the set of issues we face. I’m looking forward to the next session (and the next blog post from you) on the topic.
Doing a major site & online identity overhaul at the mo. Thanks for this, Stan.
Too bad I didn’t get to see you in Germany! Planning a trip for next month. Then may try to make it to CO in the winter so we can hang out & chill–literally.
What do the asterisks stand for (clickstream & searchstream) or you planning to flesh that out in later posts? Thank you for the educational post!!